Multidimensional work sampling was performed at a hospital-based outpatient pharmacy. Data were collected from nine full-time and five part-time pharmacists over a 45-day baseline period. Pharmacists wore silent, random-signal generators that permitted continuous work sampling. We introduced the concept of quick codes to allow pharmacists to record their work using a single letter for repetitive activities. Pharmacists recorded 4,687 observations, 90 percent using quick codes. The most common activity was checking prescriptions (36.2 percent). Detection and correction of prescribing errors was the most common reason for their work (39.4 percent). Most work-related activities were performed alone (80 percent) with little time in contact with patients or physicians. These baseline measures will be compared with future measures to assess the effect of the implementation of computerized prospective drug utilization review and clinical treatment guidelines on pharmacists' work. It is expected that these technologic and process changes will increase opportunities for pharmacists to educate patients and consult with physicians.